Medical & biological engineering & computing
-
Med Biol Eng Comput · Jan 2011
Acoustic thoracic image of crackle sounds using linear and nonlinear processing techniques.
In this study, a novel approach is proposed, the imaging of crackle sounds distribution on the thorax based on processing techniques that could contend with the detection and count of crackles; hence, the normalized fractal dimension (NFD), the univariate AR modeling combined with a supervised neural network (UAR-SNN), and the time-variant autoregressive (TVAR) model were assessed. The proposed processing schemes were tested inserting simulated crackles in normal lung sounds acquired by a multichannel system on the posterior thoracic surface. In order to evaluate the robustness of the processing schemes, different scenarios were created by manipulating the number of crackles, the type of crackles, the spatial distribution, and the signal to noise ratio (SNR) at different pulmonary regions. ⋯ Finally, the performance of the TVAR scheme was tested against a human expert using simulated and real acoustic information. We conclude that a confident image of crackle sounds distribution by crackles counting using TVAR on the thoracic surface is thoroughly possible. The crackles imaging might represent an aid to the clinical evaluation of pulmonary diseases that produce this sort of adventitious discontinuous lung sounds.
-
Med Biol Eng Comput · Jan 2011
An endotracheal intubation confirmation system based on carina image detection: a proof of concept.
In this paper, a novel system for automatic confirmation of endotracheal intubation is proposed. The system comprises a miniature CMOS sensor and electric wires attached to a rigid stylet. Video signals are continuously acquired and processed by the algorithm implemented on a PC/DSP. ⋯ As an additional validation, the system was tested using a dataset of 231 video images recorded from five human subjects during intubation. The system correctly classified 120 out of 125 non-carina images (i.e. a sensitivity of 96.0%), and 100 out of 106 carina images (i.e. a specificity 94.3%). Using a 10th-order median filter, applied on the frame-based classification results, a 100% accuracy rate was obtained.